Many studies have now demonstrated that individuals with autism attend to their visual environment in ways that differ from typically-developing peers: individuals with autism look less at people in general and less at people’s eyes in particular. While these altered patterns of looking evidence basic differences in attention to social adaptive information, few studies have directly measured how individuals, on a moment-by-moment basis, perceive the relevance of what they are looking at. In order to directly measure the perceived task relevance of visual content, during a natural viewing task, we measured the probability of blinking and blink inhibition while collecting eye-tracking data. While eye-blink data are often discarded as noise, a small research literature links blinking to visual information processing in a variety of cognitive tasks. That research demonstrates that blinking is inhibited during moments in which a viewer perceives that he or she will be processing important task-relevant visual information. Here we measure blink inhibition to index perceived stimulus relevance during natural viewing.
Objectives:
Our objectives were (1) to test whether the ongoing probability of blink inhibition during natural viewing might quantify the perception of task-relevant, salient information in typically-developing children; (2) to determine whether children with autism differed from typically-developing children in their blinking behavior; and (3) to measure whether between-group differences in timing of blink inhibition and in content of visual fixations during blink inhibition might reveal information about the perceived task-relevance of visual stimuli in children with autism relative to typically-developing peers.
Methods:
Eye-tracking data were collected from 2-year-olds with autism (n = 41) and matched, typically-developing peers (n=52) while watching movies of social interaction. Differences in physiological properties of eye-blinking were tested by measures of individual blink rate (mean blinks per minute) and expected positive correlation between age and blink rate. Movies were selected for their range in affective content, varying from low emotional valence to highly charged exchanges; affective valence was rated by 10 adult viewers naïve to experimental aims.
Results:
Children with autism and their typically-developing peers differed significantly in their timing of blink inhibition relative to scene content. Control measures—blinks per minute and correlation of blink rate with chronological age—were not significantly different. A repeated-measures ANOVA revealed a significant interaction between diagnosis and affective scene content during blink inhibition. Typically-developing toddlers inhibited their blinking during scenes with greater affective content, while blink inhibition in toddlers with autism was unrelated to affect.
Conclusions:
While physiological properties of eye-blinking were equivalent between groups, and both groups inhibited their blinking more than expected by chance, only typically-developing 2-year-olds inhibited their blinking selectively during scenes of high affective valence. These results demonstrate that patterns of blink inhibition index perceived task relevance of visual stimuli and also that they quantify reactions to, and expectations of, emotional content during natural viewing. Together these results reveal what information is missed—in the blink of an eye—by children with autism.